Boundaries don't work with OpenCL

Boundary conditions work fine on CPU and CUDA, but they don't have any effect on OpenCL. Here is a Poiseuille flow test case that shows that, just change the target in the third block. gpu_boundary.ipynb

changed the description

Maybe I have time for this on the weekend. When I submitted the PR I just stated it compiles. I guess there's still some code hidden that isn't aware of the 'opencl' target.

mentioned in merge request pystencils!148 (merged)

@kuron boundary handling was working but the SerialDataHandling.swap method not: pystencils!148 (merged)

Maybe we should add a test that checks that OpenCL and CPU yield same results in a typical LBM scenario? E.g. your example without the plotting...

Or just your notebook directly only checking convergence with analytic solution.

Thanks for tracking this down. Please do add the analytic check from my notebook as a test case for GPU, CUDA and OpenCL.

mentioned in merge request !26 (merged)

closed via merge request pystencils!148 (merged)

mentioned in commit pystencils@e87b4daf